Implement BBP RTQC (Dall'Olmo et al)

[catsch]

• aoml
• bodc
• coriolis
• csio
• csiro
• incois
• jma
• meds

[jlovellcsiro]

CSIRO does this

[Argo-JMA]

Not yet. JMA will introduce BBP RTQC by ADMT-26.

[cgrdn]

Completed for MEDS with the exception of filling BBP_ADJUSTED in RT which will be done by end of year.

[nicolagMBARI]

We are in the process of implementing this at MBARI.

We ran into some confusion about the language used for when a test is aborted. The language suggests that when a test is aborted, the test is not applied successfully. However, our understanding is that when a test is aborted, the reason why the test was aborted determines whether or not the test is applied.

For example, if there is not enough data for the test to progress, the test is aborted and we assume not applied. If the test completes all steps and the final result does not exceed a failure threshold, the test is aborted but we assume the test was successfully applied.

This distinction is not clear in the literature and impacts how we report RTQC_APPLIED in the SCI_CAL_COMMENT.

For the utility of cross-DAC analysis of these RTQC tests, we wanted to confirm how DACS were reporting RTQC_APPLIED in cases where the test is aborted.

Thank you!

[grgdll]

We are in the process of implementing this at MBARI.

We ran into some confusion about the language used for when a test is aborted. The language suggests that when a test is aborted, the test is not applied successfully. However, our understanding is that when a test is aborted, the reason why the test was aborted determines whether or not the test is applied.

For example, if there is not enough data for the test to progress, the test is aborted and we assume not applied. If the test completes all steps and the final result does not exceed a failure threshold, the test is aborted but we assume the test was successfully applied.

This distinction is not clear in the literature and impacts how we report RTQC_APPLIED in the SCI_CAL_COMMENT.

For the utility of cross-DAC analysis of these RTQC tests, we wanted to confirm how DACS were reporting RTQC_APPLIED in cases where the test is aborted.

Thank you!

Dear @catsch would you be able to please share what is being done at Coriolis?
thanks!

[catsch]

In the documentation https://doi.org/10.13155/60262
we mention two strings : RTQC_APPLIED and RTQC_FAILED to be reported in the SCI_CAL_COMMENT
RTQC_APPLIED is 1 when the test completes all steps of the test
RTQC_FAILED is 1 when the final result exceed the threshold

To correctly fill the RTQC_APPLIED, for some schemas of the documentation, you should not take into account the last "Abort" (Figure 2, Figure 4 and Figure 5 in the documentation / Figure 5, Figure 8 and Figure 14 in the Open research Europe review) to fill the RTQC_APPLIED

for example for float 3901531 (profile 125) we got RTQC_APPLIED=11111 and RTQC_FAILED=01001 (high deep value test and parking hook test failed )

I hope it helps